Remote Plasma Sputtering: Recent
Developments in Understanding the Process
S. Thornley, P. Hockley, M. Thwaites, J. Dutson
Dr James Dutson
Senior Development Engineer
Contact: james.dutson@plasmaquest.co.uk
+44(0) 1256 740680
Outline
● Plasma Quest specialise in a unique form of Sputter Deposition
● Talk will give an overview of remote plasma sputtering, and why we
use it.
–
–
–
–
–
–
Overview of Remote Plasma Sputtering
Some Empirical Benefits
Detailed Discussion of the Plasma
Implications for Sputtered Material
Large Area Application (EPPS)
Conclusions
2
Remote Plasma Sputtering
● Plasma initiation occurs in a
separate side-arm
● Electromagnets confine and
amplify plasma
● Ions are not accelerated through
a grid - different process to ion
beam
● Key to technology is the unique
properties of this plasma
3
HiTUS Configuration
● Side arm causes initiation only
followed by cascade generation
● Very high ion density
immediately next to target
● Ions have low energy, sputtering
only occurs when target is –’ve
biased
● Independent control of ion
energy and ion density
4
Plasma Bending with HiTUS
5
Some Observed Benefits
● Empirical evidence for:
– High Target Utilisation
– Controllable Stress
– Densification of Films
– Improved Adhesion
Dispersion curve for Nb2O5
Bulk
PQL
IBS
Magnetron
E-Beam
– Low Temperature Depositions
– Reactive Control
● Recent studies have focused on
explanations for these effects
6
Example Material - ITO
● Ambient process
● High adhesion onto flexible
materials eg PET, PEN, PMMA
● Properties:
Sheet Resistance < 10Ω/,
91% avg. trans.
Resistivity <4x10-4Ωcm
Flexible; < 5cm curve radius
● Main focus of study is the
Plasma
Transmission (%)
ITO deposited by HiTUS
100
80
60
PEN
Glass
40
20
0
350
550
750
950
Wavelength (nm)
7
Low Temperature Plasmas
● Plasmas consist of ions, electrons and
neutrals
● Electron-atom interaction gives:
– Ionisation – 15.8eV
– Excitation – 11.6eV (short lived)
● Equal numbers of electrons and ions,
quasi charge-neutral
http://www.iop.kiev.ua/~prtg/Index.html
● But e- have lower mass & greater
velocity which dominate properties
8
Theoretical PQL Remote Plasma
● e- are generated in an
inductively coupled skin-effect
tubular region
● e- are confined by magnetic field.
High electron density in region I
– causes excitation and is visible
● UNIFORM ion density in region
II – But low energy
● Decaying ion density in region III
(no excitation => Not visible)
Recombine at chamber walls
9
Experimental Verification
● Normal plasma: –’ve current due
to increased e- mobility
● Collector plate probe placed in
diffuse ion region (substrate) and
measured current
● I = + 1.8 mA/cm2
(nion = 6 x 1011 ions/cc)
● Collector plate plasma has
positive current due to
confinement of electrons.
10
At the Target
● No sputtering without bias so no
current without bias
● Measuring target current during
deposition (RF1.5kW, DC
>100V)
● I> 22 mA/cm2
n> 1x1013 ions/cc
● Once sputter threshold
overcome, current is
independent of Target Bias
>95%
Utility
11
Ionisation of Sputtered Species
● Collector plate current increases
with sputtering => Increased
ions arriving at substrate
● Current at plate increases with
RF power => ionisation
increases with plasma energy
● Sputtered species are ionised as
they pass through the plasma
● Beneficial for densification,
control and shows inherent iPVD
12
Substrate bias
● Ionised sputter species can be
controlled via substrate bias
● Change impact velocity, stress,
growth conditions etc.
● Some R&D systems already
have substrate bias options
available with new options for
2014.
● Works for conducting materials
(but what about insulating
materials? – see slide 17)
13
Advantages revisited
● High target use due to highly
uniform ion distribution in centre
● Ionisation of sputtered species =
increased adatom energy=
improved densification?
Dispersion curve for Nb2O5
Bulk
PQL
IBS
Magnetron
E-Beam
● No electron bombardment = less
damage
● General low energy ion
bombardment = improved
adhesion, densification & low temp
14
Large Scale Sputtering
● iEPPS designed to meet
industrial applications
● A self-contained source and
generation system extendable in
length
● V. high transfer factor as planar
target and reduced targetsubstrate separation
● BUT New geometry of plasma
15
EPPS Plasma
● Plasma is generated as a
elongated slab/oval
● Uniform plasma and electron
ring will not impinge target
● Key Questions:
– What effect will this have on
sputtering?
– What happens with insulating
targets due to lack of
electrons?
● Initial studies used to find out
16
EPPS Trials
only
● Target currents scale linearly
with plasma source power
9
8
Target Current [A]
● Initial trials limited to low powers
10
7
6
5
4
3
2
1
● Deposition rate was 20nm/min at
0
0
1
2
3
RF Power [kW]
4
30cm gap and 1.5kW power
● Extrapolating to 15cm gap and
10kW power  266nm/min
● Ion density not a problem,
>7x1012 ions/cc
17
EPPS Electron Confinement
● HiTUS collector plate biased
with +’ve and –’ve potential
Plate Current vs Plate Bias Voltage
● Currents significantly higher for
electrons due to increased
mobility
● Only small electric field required
(10-20V) to release lowest
energy electrons from magnetic
confinement
● Can use EPPS for insulating
materials
Collector Plate Current [I]
1
0
-50
-25
-1
0
25
50
-2
-3
-4
-5
-6
Plate Bias [V]
Iion= +0.25A
Ie = -5A
18
EPPS Development Programme
● EPPS has received provisional funding to commence late 2013
● PQL will be actively developing EPPS over the next 2 years
● Once EPPS established, core materials studied will be
transparent conducting oxides that are alternatives to ITO.
● In collaboration with Institute for Materials, Research and
Innovation – University of Bolton
19
Conclusions
● Remote Plasma Sputtering provides a number of key material
advantages
● The unique HiTUS plasma process confines electrons giving a
different type of plasma compared to traditional techniques
● The interaction of the plasma with substrate and sputtered species
is responsible for many advantages, including inherent iPVD
● New large area depositions are viable using EPPS technology
● New developments over the next year as these results are
exploited
20